Prosper is the first peer-to-peer lending marketplace in the United States. Since its founding in 2005, it has facilitated more than $12 billion in loans.
Prosper became SEC registered in July 2009.
New prospective borrowers were required to have a FICO 8 credit score of at least 640, while returning borrowers required 600.
Borrowers were restricted to those residing in 47 states (Iowa, Maine, North Dakota restricted).
Lenders were restricted to residents of 28 US states and the District of Columbia.
Prior to 19 December 2010, it ran as an online auction market place, where lenders and borrowers ultimately determined the loan rates, using a dutch auction-like system.
Effective 19 December 2010, Prosper used pre-set rates determined by them, evaluating each prospective borrowers credit risk.
The data set contains over 110,000 loan records and over 80 variables, from 2005 to 2014 Q1. It has information on loan dates, loan statuses, borrower’s background, loan ratings and loan rates.
Noting the changes to Prosper above, I have combined loan ratings post and pre SEC registration (July 2009) into one variable “Rating”. And have also created 2 variables to note the different time periods.
Other than those, the following variables were added:
A few questions come to mind on the topic of loans:
There’s a large range of the amounts people are looking to borrow 1k-35k. It looks like most people borrow by the thousand. Why the large bar at 4k? Is there a special rate?
Most people are taking out loans as debt consolidation, could the borrowing rates on prosper be lower than the rates outside? The other four 0 - Not available, 7 - Other, 2 - Home improvement, 3 - Business
Let’s dig a little deeper and find out more about our borrowers, also, lets take a look at the rates they are paying. Are they better on prosper, or could they not get a loan elsewhere? (Note: rates outside are about 6% - 36%)
Borrower rate ranges from 0.04-0.5, while lender yield ranges from 0-0.5. Lender yield is just a slight shift from the borrower rates, it makes sense as prosper charges a fee. Seems like the rates aren’t too far off from the norm.
Looks like most have a minimum credit score around 600-640, which is expected because of the restrictions.
Most people expect to pay use up less than 10% of their monthly income to pay for the loan. There’s some that would pay more than their income, I wonder how they are going to do that?
Most people have a debt to income ratio below 0.5, there are some exceeding 1.5, and some even exceeding 10.
When were the loans made? Did they pay their loans?
## Loan Status Summary
## ChargedOff Defaulted PastDue
## 11992 5018 2067
## Current FinalPaymentInProgress Completed
## 56576 205 38074
## Cancelled
## 5
There is a dip 2008 Q4 to 2009 Q2, post-financial crisis and possibly due to the SEC registration?
A big portion of the loans are still current, it seems like about 30% of the closed loans were charged off or defaulted.
For the loans which are considered closed, most losses were kept under 5000.
The first thing that we usually look at is the rating of a product, what does Prosper use to determine their ratings?
There are less HR in the higher income group, and also less AA in the lower income group.
Doesn’t seem like there’s too much of a relation here, it seems most have debt to income ratios below 1. The variance does seem higher for lower ratings.
There is a clear distinction here, higher credit scores do help give a higher rating.
Doesn’t seem like there’s much of a relation between rating and the proportion of monthly loan payment of monthly income.
Do people with more income need more money? Would people with their income below their monthly repayment be able to pay their loans? Since rating is one of the first things we would look at, does it determine the outcome of the loan?
Loans for those below the 100k income range seem to max out at 25k. Can see that the higher income one has, the more one can borrow.
Surprising how some expecting to pay more than their monthly income managed to complete their loan payments. Perhaps they were expecting a payout or had funds tied up for the moment.
Can see that the better ratings have less charged off and defaulted. This should be expected!
Are borrowers who get better rates more likely to pay their loans? Is a high income a factor to consider?
Lender yield vs borrow rate is pretty much to be expected. The higher the rate for the borrower, the higher yield for the lender. Looks like there are loans being completed at all levels.
Seems clear that the rating does determine a borrowers rate. There are some outliers with about 0% even at HR. Can see that for the current loans, the ratings and borrower rates are highly related.
Pre-SEC there were a lot of not-displayed income ranges. Few HR-100k+ incomes in the pre-sec period. And few AA-sub25k incomes after the SEC registration. But it doesn’t seem like the income affects how likely the loan is paid back.
Tip: You’ve done a lot of exploration and have built up an understanding of the structure of and relationships between the variables in your dataset. Here, you will select three plots from all of your previous exploration to present here as a summary of some of your most interesting findings. Make sure that you have refined your selected plots for good titling, axis labels (with units), and good aesthetic choices (e.g. color, transparency). After each plot, make sure you justify why you chose each plot by describing what it shows.
With the auction system, it looks like rates and ratings are all over! There are some HR ratings with almost 0% interest rate for the borrower. Looks like there are some really kind souls out there!
After the business model change, the borrower rate is pretty much tied in with his or her credit rating. From about 5% for an AA rating, to about 36% for a HR rating.
With the auction system, we can see the loan statuses are all over. Some which have favourable rates are also defaulting on their loans.
It seems like the lower the rates one is getting, the more likely the loan is paid off. Those with the higher rates in each rating group seems to be more likely to default.
With both the auction and fixed rates, we can see that the loan is more likely to be paid of if it has a reasonable monthly proportion and a high rating!
## Closed Loans Status Auction
## ChargedOff Defaulted PastDue
## 7598 4266 0
## Current FinalPaymentInProgress Completed
## 0 0 24506
## Cancelled
## 5
## Closed Loans Status Fixed
## ChargedOff Defaulted PastDue
## 4394 752 0
## Current FinalPaymentInProgress Completed
## 0 0 13568
## Cancelled
## 0
The change from auction to fixed rates has seemed to lower the defaulting rates. However the charged off rate seems about the same, if not higher.
The auction model benefitted the borrowers as some of them had gotten favourable rates even with their poor credit ratings.
The rates the borrowers are getting from Prosper don’t seem much different to the rates gotten outside (6% - 36%)